Last-Click vs Last-Touch vs Data-Driven: What Each Attribution Model Actually Measures

David Manela·October 6, 2026·9 min read

A row of stepping stones across still water, with only the last stone glowing.

Last-click attribution gives the final step all the credit for the whole path.

Google Ads and Google Analytics can report different conversion counts for the same campaign in the same month. Add Meta's report and the combined total can exceed the orders finance sees in the backend. Much of that gap comes from the attribution model behind each report, because each model counts a different thing.

Last-click attribution gives 100% of the credit for a conversion to the final ad or channel a customer clicked. Last-touch gives it to the final recorded interaction of any kind, which can include an ad view or a direct visit, depending on the tool. Data-driven attribution splits credit across touchpoints using a model trained on converting and non-converting paths.

All three are rules for assigning credit to conversions that already happened. None of them measures what the marketing caused. That takes an incrementality test, which we come back to at the end of this guide.

What does last-click attribution measure?

Last-click attribution gives all of the credit for a conversion to the last ad or channel the customer clicked before converting. Google Ads defines it as giving all credit to the last-clicked ad and its corresponding keyword.

Google Analytics 4 offers two versions. Paid and organic last click gives 100% of the credit to the last channel the customer clicked through, paid or organic. Google paid channels last click gives 100% of the credit to the last Google Ads channel the customer clicked, and ignores every other channel.

In GA4, direct visits get no credit under any model unless the whole path to conversion was direct. A customer who clicks a paid search ad on Tuesday and types your URL on Friday to buy is credited to paid search.

Last-click is easy to read and stable from week to week, which makes it useful for pacing budgets and spotting sudden changes.

Its blind spot is everything before the final click. It favors channels that capture demand customers already have, such as branded search and retargeting, and gives nothing to the channels that created that demand. It also cannot see an ad someone viewed without clicking. Each ad platform sees only its own clicks, so each platform's last-click report credits itself, and those reports can add up to more conversions than you actually had.

What does last-touch attribution measure, and how is it different from last-click?

Some tools and writers use the two terms interchangeably. Where a tool separates them, the difference is what counts as a touch. A click is one kind of touch. Depending on how the tool defines a touchpoint, a last-touch model may also count an ad impression or a direct visit.

Adobe Analytics uses the name Last Touch for a model that gives 100% of the credit to the touchpoint occurring most recently before conversion. Adobe's documentation says it is typically the default for any metric where no other attribution model is specified.

Meta's reporting shows how views and engagements enter the picture. In an announcement dated March 3, 2026, Meta said it was changing the definition of click-through attribution for website and in-store conversions to exclusively include link clicks. Conversions that came from a share, save or other non-link click action move to engaged-view attribution, which Meta renamed engage-through attribution. Meta's reporting can also credit a conversion to an ad that someone saw but did not click: its Marketing API includes a 1-day view attribution window, defined as 1 day after viewing the ad.

The practical difference shows up in two places:

Views and engagements. A last-touch report that counts impressions credits the final ad a customer saw, even if they never clicked it. Video and display formats gain credit under this rule.
Direct visits. A last-touch rule that counts direct visits credits direct for customers who clicked an ad earlier and came back by typing your URL. GA4's models exclude direct visits unless the whole path was direct.

Before you compare any two reports, check what each one counts as a touch. Two reports can both be labeled last-touch and still disagree, because one counts views and the other does not.

What does data-driven attribution measure?

Data-driven attribution spreads credit across the touchpoints on a conversion path, based on how much each one appears to change the chance of converting. Google Ads describes it as using your account's data to calculate the actual contribution of each interaction across the conversion path. It is the default model for most Google Ads conversion actions.

Google's GA4 documentation explains the method in two steps. The model first compares the paths of users who converted with the paths of users who did not, and uses a counterfactual approach to estimate how likely conversion was with and without each touchpoint. It then assigns fractional credit based on how much each interaction changed that estimated probability, taking into account factors such as the time between the interaction and the conversion, the ad format, and query signals.

Adobe's equivalent, called Algorithmic, allocates credit with a method based on the Harsanyi dividend from cooperative game theory.

Data-driven is now the only multi-touch model inside Google Ads and Google Analytics. Google retired the first click, linear, time decay and position-based models from both in late 2023. When it announced the change in April 2023, Google said less than 3% of Google Ads web conversions were attributed using those models. Last-click models remain available in both.

Its limits matter as much as its strengths. The model can only credit touchpoints the tool recorded, so a podcast ad or a TV spot gets nothing. It estimates contribution from patterns in observed paths and does not run an experiment. Inside an ad platform, the company selling the media also builds the model that grades it, and you cannot inspect how that model works.

How do the three models compare side by side?

Last-click: credits the final click. Suited to pacing and week-over-week monitoring. Misses views and every touch before the last click.
Last-touch: credits the final recorded interaction, which may be a view or a visit. Suited to seeing which exposure came last. Results depend on what the tool counts as a touch.
Data-driven: splits credit using a model of converting and non-converting paths. Suited to cross-channel credit and to bidding inside the platform that built it. Cannot credit untracked media and does not prove cause.
Incrementality test, for contrast: measures the conversions that would not have happened without the spend. Suited to deciding whether to move budget. Slower and more costly to run than reading a report.

Why do attribution models disagree with each other and with finance?

Reports disagree for reasons that have little to do with which model is smarter. The common causes:

Different touch definitions. One report counts clicks only. Another adds views or engagements.
Different lookback windows. In GA4, the default window is 30 days for acquisition key events (first_open and first_visit), with 7 days as the alternative, and 90 days for all other key events, with 30 or 60 days as alternatives. Separately, Google announced on August 11, 2026 that Google Analytics conversions now support custom lookback windows of any whole number of days from 1 to 90 for click-through conversions and from 1 to 30 for engaged-view conversions, configured under Advertising > Conversion management in Google Analytics or in the linked Google Ads account.
Different scope. Google Ads only reports conversions that followed an interaction with that Google Ads account. A cross-channel tool such as GA4 may credit the same sale to another channel.
Different dates. Google notes that Google Ads reports conversions on the ad impression date, while other reporting tools attribute them to the conversion date. A campaign that runs late in the month can show its conversions in different months in different reports.
Different handling of direct traffic. GA4 excludes direct visits from credit. Other tools may not.

An illustration with made-up numbers shows how one sale becomes several. A customer clicks a Meta ad on Monday, searches your brand name and clicks a Google ad on Thursday, then types your URL on Saturday and places a $120 order. Meta's report can claim the order as a click-through conversion. Google Ads claims it under last click. GA4's paid and organic last click credits Google paid search, because the direct visit is excluded. A last-touch rule that counts direct visits credits direct. Finance sees one $120 order. The two ad platform reports together show $240.

This double counting is one of the reasons strong ROAS does not always show up in the P&L.

Which attribution model should you use for which decision?

Pick the model by the decision in front of you. We cover that approach in more depth in Stop Debating Attribution Models. Use Them All. A practical split:

Daily and weekly monitoring: last-click, held constant, so the trend reflects performance alone.
Bidding and optimization inside Google Ads: data-driven, which is the default for most conversion actions.
Understanding the role of video and display: data-driven, or a last-touch report that includes views, read as a directional signal.
Moving budget between channels: none of the three on its own. Use an incrementality test or a marketing mix model calibrated with experiments. Our guide to measuring incrementality without a holdout group covers the options.

Exactius operators have analyzed more than $848M in marketing spend.

How do you set up attribution reporting that finance will trust?

Choose one source of record. Report revenue to finance from your own order or billing system, on contribution margin where you can, and treat platform numbers as claims to reconcile against it.
Document every report's rules. For each tool, record the model, what counts as a touch, the lookback window and how direct traffic is treated.
Hold the rules constant. Note the date of any change to a model or window, because it breaks comparison with earlier periods.
Reconcile every month. Compare the sum of platform-reported conversions with backend orders. The ratio is a rough gauge of double counting. It cannot tell you which platform is overstating, and it shifts whenever the channel mix changes.
Calibrate with experiments. Run incrementality tests on your largest channels and use the results to adjust how much weight each platform's attributed numbers get.

Sources

If your platform reports and your backend never agree, book a call and an Exactius operator will map what each report counts and where the gap comes from.

Exactius is a full-funnel growth agency accountable for its clients' P&L. Its AI-enabled senior operators provide performance marketing, strategy, creative, and whole-business analytics and data science, engaged one function at a time or as a full team. It serves consumer and B2B companies where paid marketing is a main growth lever, through two practices: one for companies from $5M to $100M and one for companies from $100M to $1B.

Tags:attribution modelslast-click attributionlast-touch attributiondata-driven attributionGA4 attributionGoogle Ads attributionMeta attributionmarketing measurement

David Manela

David Manela is the founder of Exactius and creator of the Growth Operating System — a framework for deploying capital-efficient, compounding growth inside scaling companies.

FAQ

Frequently asked

What is the difference between last-click and last-touch attribution?

Both give 100% of the credit to a single final interaction. Last-click counts only clicks. Last-touch counts the final recorded interaction of any kind, which, depending on the tool, can include an ad view or a direct visit. Some tools use the two terms interchangeably, so check each tool's definition before comparing reports.

Which attribution models are available in Google Analytics 4?

GA4 offers data-driven attribution and two last-click models: paid and organic last click, and Google paid channels last click. Google retired the first click, linear, time decay and position-based models in late 2023. In every GA4 model, direct visits receive no credit unless the entire path to conversion was direct.

How does data-driven attribution work?

Google's version compares the paths of users who converted with the paths of users who did not, estimates how much each touchpoint changed the probability of converting, and assigns fractional credit on that basis. It weighs factors such as the time between the interaction and the conversion and the ad format. It can only credit touchpoints the tool recorded.

Does data-driven attribution measure incrementality?

No. Data-driven attribution divides credit for conversions that already happened, based on patterns in observed paths. Incrementality measures how many conversions would not have happened without the spend, which requires an experiment or a model calibrated with one. A channel can earn a large share of data-driven credit and still drive few incremental sales.

Why do Google Ads and Google Analytics report different conversion numbers?

They can use different attribution models, lookback windows and scopes. Google Ads only reports conversions that followed an interaction with that Google Ads account, while GA4 credits all channels. Google also notes that Google Ads reports conversions on the ad impression date, while other reporting tools attribute them to the conversion date.

Which attribution model should a growing company use?

Use more than one, each for the decision it suits. Last-click works for daily and weekly monitoring when it is held constant. Data-driven works for cross-channel credit and for bidding inside Google Ads. For decisions about moving budget between channels, validate with incrementality tests or a marketing mix model calibrated with experiments.

What attribution lookback window should I use?

Start from your typical time between first click and purchase. GA4 defaults to 30 days for acquisition key events and 90 days for other key events. Separately, since August 11, 2026, Google Analytics conversions managed under Advertising > Conversion management support custom click-through windows from 1 to 90 days. A window shorter than your purchase cycle drops real conversions. A much longer one can credit ads that had little to do with the sale.

Keep going

Ready to fix the system?

Your growth system is either compounding or degrading.

Book a diagnostic call. We'll identify where your growth system is breaking and what it's costing you.

Book a call ← More articles